3.1.1 Yeast two-hybrid library screening
In order to determine PHF5a binding partners the yeast-two hybrid library screening was performed. The Matchmaker murine embryo 11.5-days library (BD Clontech) was screened by using the PHF5a protein as a bait. The BK/PHFa construct used for screening of the yeast two-hybrid library was generated by fusing the complete ORF of PHF5a (Accession No.: NM_026737) in frame into the pGBKT7 vector (Figure 4)
Figure 4. A. Schematic representation of the yeast two-hybrid bait construct BK/PHFA. The PHF5a fragment was amplified with primers pRT-21 FP and pRT-21 RP, purified, digested with EcoRI and cloned in frame to EcoRI site of pGBKT7 vector. The correct orientation was determined by double test digestions of resulting DNA clones with SstI/PstI. PADH1 -truncated S. cerevisiae ADH1 promoter, GAL4 DNA-BD- GAL4 DNA binding domain, (PT7)- T7 RNA polymerase promoter, (PT7 ADH1)- transcription termination signals, (MCS)-Multiple Cloning Site, (pUC ori)- pUC plasmid replication origin, (Kanr)- Kanamycin resistance gene, (2 µ
ori)-PHF5a 330 bp EcoRI
EcoRI SstI PstI
NH2 c-Myc finger1 finger2 finger3 L-zipper COOH
PHF5a GAL4 BD
A
B
GAL4 BD
Yeast 2 µ replication origin, (TRP1)- TRP1 coding sequences, (f1 ori)- f1 bacteriophage origin of replication.
Circular pGBKT7 vector map was adapted from pGBKT7 Vector Information, Protocol # PT3248-5, CLONTECH Laboratories, Inc. (1999), B. Schematic representation of the PHF5a GAL4 DNA-BD fusion protein. (GAL4 BD)- GAL4 DNA binding domain, (c-Myc)- c-Myc epitope tag, (finger1)- zinc finger, (finger2)-zinc finger 2, (l-zipper)- leucine zipper.
Additional tests were performed to exclude possible autoactivation of the reporter genes by the PHF5a-BD fusion protein alone. The Saccharomyces cerevisiae strain AH109 was transformed with the bait construct BK/PHF5a and served as a mating partner for the host strain Y187 pretransformed with the Matchmaker 11.5-days mouse embryo cDNA library.
Both strains were combined and incubated for 24 hours before plating on the medium stringency selective SD plated (-Leu, -Try, - His). Following selection on medium stringency conditions, 200 resulting clones were tested on the SD medium (-Leu , -Try, -His, -Ade) and supplemented in X α-Gal (Figure 5). Approximately 80% of clones growing on high stringency plates turned blue as a result of MEL1 reporter activation. Its product _-galactosidase is a secreted enzyme and could by assayed directly on selective SD plates.
About 100 positive clones (robust > 2 mm blue colonies) were picked and tested for the presence of a cDNA library insert by the colony PCR using pACT2 vector-specific primers Y2H2 and Y2HAD3’. Subsequently, DNA sequencing of 60 of these clones revealed that cDNAs from four genes were present multiple times as independent overlapping clones, namely U2AF35, SRp40, Ddx1 and mDomino.
Figure 5. Schematic representation of the strategy used for the selection of PHF5a interacting proteins by the yeast two-hybrid embryo 11.5-days library screening. Picture partially adopted form the BD Matchmaker Pretransformed Libraries Users Manual, BD Biosciences Clontech, (1999), modified.
The following cDNA clones were selected for further work: clone no.35 containing the cDNA fragment of the splicing factor U2AF35 from nucleotide position 21-740 (Accession No.:NM_024187) corresponding to the complete ORF of U2AF35; no. 94 with the shortest interacting fragment of the splicing factor SRp40 from nucleotide position 659-832 (Accession No.:NM_009159) spanning of the C-terminal region rich in arginine-serine (RS domain); no. 91 with the shortest interacting cDNA fragment of the ATP-dependent RNA helicase DEAD/H (Ddx1) from nucleotide position 1660- 2303 (Accession No.:NM_134040)
pGBKT7 pACT2
cDNA inserts cDNA PHF5a
-TRP
-TRP, -LEU, -HIS
-TRP, -LEU, -HIS, -ADE, + X α-GAL Transformation into
AH109 (MATa) strain
2X YPDA
Pretransformed Into Y187 (MATα) strain
Mating
Medium stringency selection
High stringency selection
and no. 100 containing the shortest interacting cDNA fragment of the SWI2/SNF2-type ATPase/helicase protein mDomino-s from nucleotide position 1880-2516 (Accession No.:
AB092695). The constructs: AD/SRp40-RS, AD/U2AF35, AD/mDomino, and AD/Ddx1 were prepared by digestion with NcoI/SfiI of colony PCR products obtained from yeast colonies and cloning in frame into the pGADT7 vector (Figure 6).
U2AF 719bp Ddx1 643bp mDomino 636bp SfiI
NcoI
SRp40 173bp
Y2AD3’
Y2H2FP
NH2 COOH
U2AF35 clone no. 14 U2AF35 clone no. 16 U2AF35 clone no. 2
A
B
U2AF35 clone no. 35
U2AF35
Figure 6. A. Schematic representation of pGADT7 constructs containing cDNA fragments isolated by the yeast two-hybrid assay. The yeast colony-PCR products generated with Y2H2 FP and Y2AD3’ primers from the yeast clone 35 (U2AF35), clone 91(Ddx1), clone 100 (mDomino) and clone 94 (SRp40) were purified, double digested with enzymes NcoI/SfiI and cloned in frame into the pGADT7 vector. (PADH1)- Full-length S. cerevisiae ADH1 promoter, (SV40 NLS GAL4 AD)- GAL4 acceptor domain polypeptide with SV40 Nuclear Localization Signal, (PT7)- T7 RNA polymerase promoter, (HA)- HA epitope tag, (MCS)- Multiple Cloning Sites, (TADH1 )-Transcription termination signal, (LEU2)- LEU2 coding sequences, (pUC ori)- pUC plasmid replication origin, (Ampr)- Ampicillin resistance gene, (2 µ ori)- Yeast 2 µ replication origin. Circular pADT7 vector map adapted from pGADT7 Vector Information, Protocol # PT3249-5, CLONTECH Laboratories, Inc., 1999. B. Schematic representations of domain structures of U2AF35, SRp40, Ddx1 and mDomino with annotated corresponding cDNA clones found in the yeast two-hybrid library screening (according to the NCBI Conserved Domain Search v. 2.00). (RRM)- RNA recognition motif, (SrmB)- Superfamily II DNA and RNA helicases, (DEADc)- DEAD-box helicases, (SPRY)- Domain in SPla and the RYanodine Receptor, (HELICc)- Helicase superfamily
terminal domain, (DEXH)- DEXH-box helicases, (SNF2_N)- SNF2 family N-terminal domain, (HepA)-Superfamily II DNA/RNA helicases
Three of the identified interaction partners (U2AF35, SRp40 and Ddx1) consist of either domains characteristic for RNA processing proteins or were previously described to be involved in this process (Zhang et al., 1992; Screaton et al., 1995; Bléoo et al., 2001). Non-snRNPprotein splicing factors SRp40 and U2AF35 contain domainsrich in alternating serine and arginine residues (RS domains). Clone no. 94 encoding the C-terminal RS domain of the SRp40 protein was sufficient to maintain the interaction with the PHF5a protein, therefore, construct encoding the RS domain of U2AF35 was generated. For this purpose U2AF35 cDNA fragment from nucleotide position 540-740 was amplified by RT-PCR on total RNA extracted from murine 11.5-days embryos using primers U2AF-RS FR, U2AF-RS RP and cloned in frame into the pGADT7 vector.
Furthermore, yeast cotransformants with the PHF5a protein as a pray and the RS domain of U2AF35 as a bait could grow on selective SD plates (-Leu, -Try, -His, -Ade). This result suggests that PHF5a would act as an arginine-serine rich (RS) domain binding protein.
Ddx1 and mDomino are ATP-dependent helicases and Ddx1 is believed to be involved in RNA processing (Godbout et al., 1993; Bléoo et al., 2001)
3.1.2 Determination of PHF5a binding affinity to different RS domains by the quantitative _-galactosidase assay
Double transformants of the bait construct and constructs containing library founders resulted in blue colonies as a result of MEL1 activation. The product of the MEL1 gene, namely α-galactosidase, could be quantitatively measured in the yeast liquid medium and therefore, it allows for directly comparing the relative strength of PHF5a interactions with different proteins.
In order to answer the question, whether PHF5a interacts unspecifically with all RS domain-containing proteins or if PHF5a binding is rather restricted to proteins found in the yeast two-hybrid library screening, fragments containing RS domains of three other SR proteins were cloned and tested. For that purpose, constructs encoding RS domains of SRp20 (Accession No.: NM_013663) containing a cDNA fragment from nucleotide position 355-612; SRp30c (Accession No.:NM_025573) containing a cDNA fragment from nucleotide position 406-801; ASF/SF2 (Accession No.: NM_173374) containing a cDNA fragment from nucleotide position 697-882 and U2AF35 containing a cDNA fragment from nucleotide position 540-740 were generated by RT-PCR, and cloned in frame into pGADT7 vector (Figure 7).
SRp30 RS 395bp SRp20 RS 257bp ASF/SF2 RS 185bp
EcoRI
U2AFRS 200bp BamHI
A
B
Figure 7. A. Schematic representation of pGADT7 constructs containing RS domains of splicing factors SRp30, SRp20, ASF/SF2 and U2AF35 used in the quantitative _-galactosidase assay. cDNA fragments corresponding to the RS domain of SRp30, SRp20, ASF/SF2 and U2AF35 were amplified by the one-step RT-PCR using primers RS-SRp30 FP, RS-SRp30 RP; RS-SRp20 FP, RS-SRp20 RP; RS-ASF/SF2 FP, RS-ASF/SF2 RP and RS-U2AF FP, RS-U2AF RP, respectively. RT-PCR products were purified, digested with EcoRI/BamHI and cloned in frame into the pGADT7 vector. . B. Schematic representation of domain structure of proteins containing RS domains namely: SRp30, SRp20, ASF/SF2, U2AF35 and SRp40. Localisations of the corresponding pGADT7 clones containing RS domains are annotated as grey lines. Domain architecture according to the NCBI Conserved Domain Search v. 2.00.
The BK/PHFa bait construct was cotransformed together with constructs containing RS domains of following splicing factors: ASF/SF2, SRp30 and SRp20. Subsequently, yeast cotransformants were tested on selective SD plates (-Leu, -Try, -His, -Ade). Yeast cells containing the bait PHF5a-BD construct and the SRp30-AD construct could not grow on these plates. Cotransformants bearing the bait PHF5a-BD fusion protein and ASF/SF2-AD or SRp20-AD fusion constructs could form colonies, but they did not turn blue in the presence of x-α Gal.
The quantitative α-galactosidase assay was implemented to compare the strength of PHF5a interaction with different RS domains (Figure 8). The galactosidase activities in the yeast clones bearing the PHF5a bait and RS domains of ASF/SF2 and SRp20 were almost undetectable, whereas the strongest activity was measured for the DEAD/H (Ddx1) helicase
and U2AF35. Taken together, these results indicate that protein interactions between PHF5a and ASF/SF2, SRp20, SRp30 are either very weak or are not detectable in this assay.
Figure 8. The α-Gal quantitative assay shows the relative strength of interactions between PHF5a and splicing proteins. The bars show the fold induction of α-galactosidase activity as compared to empty vectors.
3.1.3 Coimmunoprecipitation of the PHF5a protein with interaction partners
The coimmunoprecipitation assay was performed to answer the question whether in vitro protein-protein interactions would corroborate the interactions found by the yeast two-hybrid assay. For this experiment, pGADT7 constructs coding hemagglutinin (HA) tagged versions of SRp40, U2AF35, Ddx1 and mDomino were used. Recombinant [35S] labelled proteins were produced by the in vitro transcription-translation experiment as described in Materials and Methods. These proteins and their putative binding partner, i.e. the c-Myc tagged PHF5a protein were used in the immunoprecipitation assay.
After 1 hour incubation at room temperature the c- Myc tagged PHF5a protein and all four proteins found by the yeast two-hybrid assay and expressed as a fusion protein with a HA tag could be efficiently coimmunopreciptated. This interaction was detectable in all cases using both combinations, i.e. by using either c-Myc or HA antibodies recognizing adequate epitopes. There was no unspecific precipitation of the HA tagged proteins by using the c-Myc antibody and the c-Myc tagged proteins by using the HA antibody (Figure 9 and Figure 10).
Figure 9. A. SDS-PAGE analysis showing that the murine PHF5a protein coimmunoprecipitates with the U2AF35 protein. The proteins were in vitro transcribed/translated from the constructs BK/PHF5a (c-Myc epitope) and AD/U2AF35 (HA epitope). PHF5a-c-Myc and U2AF35-HA radioactively 35S labeled fusion products were generated. After immunoprecipitation and elution from the Protein A beads, 10µl of the immunoprecipitate was loaded onto 4-12% SDS polyacrylamide gel. Lane 1: U2AF35+ HA antibody; Lane 2: PHF5a + c-Myc antibody (note unspecific band at ~38 kDa); Lane 3: U2AF35+ PHF5a + HA antbody; Lane 4: U2AF35 + PHF5a
antibody
+ c-Myc antibody; Lane 5: negative control U2AF35+ c-Myc; Lane 6: negative control PHF5a + HA antibody.
B. Schematic view of recombinant fusion proteins used in the coimmunoprecipitation assay.
Figure 10. A. SDS-PAGE analysis showing that the murine PHF5a protein coimmunoprecipitates with Ddx1, SRp40 and mDomino-s. The 35S labeled proteins with a HA epitope were produced from the constructs AD/Ddx1-RS, AD/SRp40-RS, and AD/mDomino-RS, respectively. After immunoprecipitation and elution from Protein A beads, 10 µl of the immunoprecipitate was loaded onto 4-12% SDS polyacrylamide gel. Lane 1:
PHF5a + c-Myc antibody; Lane 2: Ddx1-RS + HA anibody; Lane 3: SRp40-RS + HA antibody; Lane 4:
mDomino-RS + HA antibody; Lane 5: PHF5a + Ddx1-RS + HA antibody; Lane 6: PHF5a + Ddx1-RS + c-Myc
PHF5a Ddx1 SRp40 mDomino PHF5a/Ddx1 PHF5a/Ddx1 PHFa/SRp40 PHFa/SRp40 PHF5a/ mDomino PHF5a/ mDomino PHF5a Ddx1 SRp40 mDomino-s
1 2 3 4 5 6 7 8 9 10 11 12 13 14
PHF5a + mDomino-RS + HA antibody; Lane 10: PHF5a + mDomino-RS + c-Myc antibody; Lane 11: negative control PHF5a + HA antibody; Lane 12: negative control Ddx1-RS + c-Myc antibody; Lane 13: negative control SRp40-RS + c-Myc antibody; Lane 14: negative control mDomino-RS + c-Myc antibody. B. Schematic view of recombinant fusion proteins used in the coimmunoprecipitation assay.
In order to determine the binding specificity of PHF5a and U2AF35, coimmunoprecipitation assay in buffers containing different ionic strength was performed.
The complex of the PHF5a protein with U2AF35 was even stable after extensive washing in a wide range of increasing NaCl concentrations in the washing buffer, suggesting a rather hydrophobic than ionic interaction between the PHF5a protein and the RS domain containing protein U2AF35 (Figure 11).
Figure 11. SDS-PAGE analysis demonstrating that the interaction between the PHF5a protein and the U2AF35 protein examined by the immunoprecipitation assay is salt-independent. Lane 1: PHF5a-c-Myc protein immunoprecipitated with PHF5a antibody; Lane 2- 8: PHF5a + U2AF35 + c-Myc antibody together with the increasing gradient of NaCl in the washing buffer.
U2AF35 PHF5a
1 2 3 4 5 6 7 8
150 150 200 250 300 400 500 1000